  

	******************************************************************
	**
	**
	**		NAME:		RACHEL BRULE & NIKHAR GAIKWAD 
	**		DATE: 		March 31, 2020
	**		PROJECT: 	Culture, Capital and the Political Economy Gender Gap
	**
	**		DETAILS: 	This file analyzes the survey data collected by the co-	
	**					authors (Brule and Nikhar).
	**					Secondary data analysis drawn from sources listed below.
	**
	**		ORIGINAL		
	**		Version: 	Stata SE 14
	**		
	**		LATEST
	**		Version:	Stata MP 16
	**
	******************************************************************
	
	




clear
clear mata
clear matrix


global input_path "~/Replication/JOP_RawData"
global output_path "~/Replication/JOP_Output"
global analysis_datasets = "~/Replication/JOP_CleanData"


*global input_path "/Users/rebrule/Dropbox/MatrilinealStudy/Replication/JOP_RawData"
*global output_path "/Users/rebrule/Dropbox/MatrilinealStudy/Replication/JOP_Output"
*global analysis_datasets = "/Users/rebrule/Dropbox/MatrilinealStudy/Replication/JOP_CleanData"


cd "$analysis_datasets"



********************************************************************************
********* ANALYSIS OF SECONDARY DATA *******************************************
********************************************************************************


********************************************************************************
********* FIGURE 1 & APPENDIX TABLE A1 *****************************************
********************************************************************************


****************
** ISSP 2006 ***
****************

use "ISSP_2006_truncated.dta", clear

* Source: ISSP 2006 - Role of Government IV, Basic Questionnaire


* STEP 1. Create Gender Dummies *
*--------------------------------

tostring sex, gen(sex_code)
*Note: male = 1, female = 2, 72 observations = .
 
gen male=1 if sex==1
replace male=0 if male==. 
 
gen female=1 if sex==2
replace female=0 if female==.



* STEP 2. Descriptive Stats *
*----------------------------


***Interest in Politics & Political Knoweldge***
*-----------------------------------------------

**Q10: How interested would you say you personally are in politics? 

tab V44

tostring V44, gen(v44)
destring v44, replace 

tab v44
*scores range from 1 (very interested) to 5 (not at all interested)

*Flip ranking order of answer choices (for ease of interpreation)
gen polint = .
replace polint = 1 if v44 == 5
replace polint = 2 if v44 == 4
replace polint = 3 if v44 == 3
replace polint = 4 if v44 == 2
replace polint = 5 if v44 == 1
 
tab v44 polint
*now scores vary from 1 "Not at all interested" to 5 "Very interested" 

*ttest v44, by(sex)
ttest polint, by(sex)
*Results ((highly significant)):  t = 32.3666   p-value= 0  
*men have significantly higher mean (i.e. are more interested in politics)


**Q11c: "I feel that I have a pretty good understanding of the important
**       political issues facing our country"

tab V47

tostring V47, gen(v47)
destring v47, replace 

tab v47
*scores range from 1 (strongly agree) to 5 (strongly disagree)

*Flip ranking order of answer choices (for ease of interpretation) 
gen polund = .
replace polund = 1 if v47 == 5
replace polund = 2 if v47 == 4
replace polund = 3 if v47 == 3
replace polund = 4 if v47 == 2
replace polund = 5 if v47 == 1
 
tab v47 polund
* now scores range from 1 "Strongly disagree" to 5 "Strongly agree"

*ttest v47, by(sex)
ttest polund, by(sex)
*Results ((highly significant)):  t = 34.9693   p-value= 0  
*men have significantly higher mean (i.e. claim more understanding of politics)



***Trust and Accountability***
*----------------------------

**Q11a: "People like me don't have any say about what the government does"

tab V45

tostring V45, gen(v45)
destring v45, replace

tab v45
*scores range from 1 (strongly agree) to 5 (strongly disagree)

*Flip ranking order of answer choices (for ease of interpretation) 
gen polsay = .
replace polsay = 1 if v45 == 5
replace polsay = 2 if v45 == 4
replace polsay = 3 if v45 == 3
replace polsay = 4 if v45 == 2
replace polsay = 5 if v45 == 1

tab v45 polsay
*now scores range from 1 "Strongly disagree" to 5 "Strongly agree"

*ttest v45, by(sex)
ttest polsay, by(sex)
*Results ((highly significant)):  t =  -7.0028   p-value= 0  
*men have significantly lower mean 
*(i.e. disagree more with the statement that people have say in gov. actions) 



***Government Responsibility: Equality***
*----------------------------------------

**Q7g: On the whole, do you think it should or should not be the government's 
**   responsibility to...reduce income differences between the rich and the poor

tab V31

tostring V31, gen(v31)
destring v31, replace

tab v31
*scores range from 1 (definitely should be) to 4 (definitely should not be)

*Flip ranking order of answer choices (for ease of interpretation) 
gen govinc = .
replace govinc = 1 if v31 == 4
replace govinc = 2 if v31 == 3
replace govinc = 3 if v31 == 2
replace govinc = 4 if v31 == 1

 
tab v31 govinc
*now scores range from 1 "Definitely should not be" to 5 "Definitely should be"

ttest v31, by(sex)
ttest govinc, by(sex)
*Results ((highly significant)):  t = -11.2930  p-value = 0
*women have significantly higher mean 
*(i.e. agree more with gov. responsibility to reduce income differences)


***Government Responsibility: Jobs***
*------------------------------------

**Q7a: On the whole, do you think it should or should not be the government's 
**     responsibility to ... provide a job for everyone who wants one

tab V25

tostring V25, gen(v25)
destring v25, replace

tab v25
*scores range from 1 (definitely should be) to 4 (definitely should not be)

*Flip ranking order of answer choices (for ease of interpretation) 
gen govjob = .
replace govjob = 1 if v25 == 4
replace govjob = 2 if v25 == 3
replace govjob = 3 if v25 == 2
replace govjob = 4 if v25 == 1

tab v25 govjob
*now scores range from 1 "Definitely should not be" to 4 "Definitely should be"


ttest v25, by(sex)
ttest govjob, by(sex)
*Results ((highly significant)):  t = -15.7817   p-value = 0 
*women have significantly higher mean 
*(i.e. agree more with gov. responsibility to provide jobs)


***Government Financing***
*-------------------------

**Q5f: Here are some things the government might do for the economy. 
**     Please show which actions you are in favour of and which you are against:
**     Reducing the working week to create more jobs

tab V16

tostring V16, gen(v16)
destring v16, replace

tab v16
*scores range from 1 (strongly in favor) to 5 (strongly against)

*Flip ranking order of answer choices (for ease of interpretation) 
gen govjob2     = .
replace govjob2 = 1 if v16 == 5
replace govjob2 = 2 if v16 == 4
replace govjob2 = 3 if v16 == 3
replace govjob2 = 4 if v16 == 2
replace govjob2 = 5 if v16 == 1

tab v16 govjob2
* now scores range from 1 "Strongly against" to 5 "Strongly in favor"

ttest v16, by(sex)
ttest govjob2, by(sex)
*Results ((highly significant)): t =  -15.1182  Pr(|T| > |t|) = 0
*men have significantly lower mean 
*(i.e. disagree more with gov. financing Projects for new jobs)


**Q6b:Listed below are various areas of government spending.
**    Please show whether you would like to see more or less government spending 
**    in each area. Remember that if you say "much more", it might require
**    a tax increase to pay for it: Health

tab V18

tostring V18, gen(v18)
destring v18, replace

tab v18
*scores range from 1 (spend much more) to 5 (spend much less)

*Flip ranking order of answer choices (for ease of interpretation) 
gen govheal     = .
replace govheal = 1 if v18 == 5
replace govheal = 2 if v18 == 4
replace govheal = 3 if v18 == 3
replace govheal = 4 if v18 == 2
replace govheal = 5 if v18 == 1

tab v18 govheal
* now scores range from 1 "Spend much less" to 5 "Spend much more"

ttest v18, by(sex)
ttest govheal, by(sex)
*Results ((highly significant)): t =  -12.4307  Pr(|T| > |t|) = 0
*women have significantly higher mean 
*(i.e. agree more with gov. spending health)



**Q6f:Listed below are various areas of government spending.
**    Please show whether you would like to see more or less government spending 
**    in each area. Remember that if you say "much more", it might require
**    a tax increase to pay for it: Retirement

tab V22

tostring V22, gen(v22)
destring v22, replace

tab v22
*scores range from 1 (spend much more) to 5 (spend much less)

*Flip ranking order of answer choices (for ease of interpretation) 
gen govret     = .
replace govret = 1 if v22 == 5
replace govret = 2 if v22 == 4
replace govret = 3 if v22 == 3
replace govret = 4 if v22 == 2
replace govret = 5 if v22 == 1
 
tab v22 govret
* now scores range from 1 "Spend much less" to 5 "Spend much more"

ttest v22, by(sex)
ttest govret, by(sex)
*Results ((highly significant)): t =  -14.4268  Pr(|T| > |t|) = 0
*women have significantly higher mean 
*(i.e. agree more with gov. spending on retirement)


**Q6g:Listed below are various areas of government spending.
**    Please show whether you would like to see more or less government spending 
**    in each area. Remember that if you say "much more", it might require
**    a tax increase to pay for it: Unemployment benefits

tab V23

tostring V23, gen(v23)
destring v23, replace

tab v23
*scores range from 1 (spend much more) to 5 (spend much less)

*Flip ranking order of answer choices (for ease of interpretation) 
gen govunem     = .
replace govunem = 1 if v23 ==5
replace govunem = 2 if v23 ==4
replace govunem = 3 if v23 ==3
replace govunem = 4 if v23 ==2
replace govunem = 5 if v23 ==1
 
tab v23 govunem
* now scores range from 1 "Spend much less" to 5 "Spend much more"

ttest v23, by(sex)
ttest govunem, by(sex)
*Results ((highly significant)): t =  -11.9480  Pr(|T| > |t|) = 0
*women have significantly higher mean 
*(i.e. agree more with gov. spending on unempl. benefits)

save "ISSP_2006.dta", replace


**********
** WVS ***
**********

cd "$analysis_datasets"

use "WVS_2016_truncated.dta", clear

* Source: WVS 2010-2014 round


* STEP 1. Create Gender dummies/ Drop "no answer" *
*---------------------------------------------------

tab V240

tostring V240, gen(sex)
destring sex, replace
* Note: male = 1, female = 2

tab sex

*Treat "Unknown" and "No answer" as missing values 
replace sex = . if V240 == -5
replace sex = . if V240 == -2


* STEP 2. Name & Classify Countries *
*------------------------------------

**NOTE: must install sdecode.pkg first
ssc install sdecode.pkg


sdecode V2, gen(country_name)

tab country_name



* STEP 3. T-tests *
*------------------

***Government Spending***
*------------------------

**Q131. Many things are desirable, but not all of them are essential 
**      characteristics of democracy. Please tell me for each of the following
**      things how essential you think it is as a characteristic of democracy. 
**      Use this scale where 1 means “not at all an essential characteristic 
**      of democracy” and 10 means it definitely is “an essential characteristic 
**      of democracy”: Governments tax the rich and subsidize the poor.

tab V131

tostring V131, gen(v131)
destring v131, replace

tab v131

 *Treat "Inappropriate", "No answer" and "Don´t know" as missing values
 
replace v131=. if V131==-5
replace v131=. if V131==-2
replace v131=. if V131==-1

tab v131
*scores vary from 1 "Not an essential characteristic of democracy" to 
* 10 "An essential characteristic of democracy"
*(i.e. higher mean value indicates more preferences for tax and subsidies)


* For World (in main table)
ttest v131, by(sex)
*Results ((highly significant)): t =  -4.0101  Pr(|T| > |t|) = 0.0001 
*women have higher mean 
*(i.e. have higher preferences for tax and subsidies)


* For India (in main table)
ttest v131 if country_name=="India", by(sex)
*Results (significant): t =  -1.6933  Pr(|T| > |t|) = 0.0905
*women have higher mean 
*(i.e. have higher preferences for tax and subsidies)


***Government Responsibility***
*------------------------------

*V98. Government should take more responsibility to ensure that everyone issues
*	  provided for = 1; People should take more responsibility to provide force
*	  themselves = 10.
*(i.e. lower mean value indicates perception of more government responsibility) 

tab V98

tostring V98, gen(v98)
destring v98, replace

tab v98

 *Treat "Inappropriate", "No answer" and "Don´t know" as missing values
 
replace v98=. if V98==-5
replace v98=. if V98==-2
replace v98=. if V98==-1

tab v98

 *Flip ranking order of answer choice (in order to facilitate result interpretaion)
 replace v98=1 if V98==10
 replace v98=2 if V98==9
 replace v98=3 if V98==8
 replace v98=4 if V98==7
 replace v98=5 if V98==6
 replace v98=6 if V98==5
 replace v98=7 if V98==4
 replace v98=8 if V98==3
 replace v98=9 if V98==2
 replace v98=10 if V98==1
  
tab v98 V98
*now scores vary from 1 "People should take more responsibility..." to 10 "The 
*government should take more responsibility..."

*FOR WORLD:
ttest v98, by(sex)
*Results ((highly significant)): t =   -5.2278  Pr(|T| > |t|) = 0
*women have higher mean (i.e. agree more that the gov should provide for everyone)

*FOR INDIA:
ttest v98 if country_name=="India", by(sex)
*Results ((highly significant)): t =  -3.4448   Pr(|T| > |t|) = 0.0006
*women have higher mean (i.e. agree more that the gov should provide for everyone)



***Wealth Ownership***
*----------------------

*If a woman earns more money than her husband, it's almost certain to cause 
*problems

tab V47

tostring V47, gen(v47)
destring v47, replace

tab v47

 *Treat "Inappropriate", "No answer" and "Don´t know" as missing values
 
replace v47=. if V47 == -5
replace v47=. if V47 == -2
replace v47=. if V47 == -1

tab v47
*scores vary from 1 "Agree" to 3 "Disagree"
*(i.e. lower mean value indicates higher consensus on the statement)

*Flip ranking order of answer choice (to facilitate interpretation)
 replace v47=3 if V47==1
 replace v47=1 if V47==3
 
* For World
ttest v47, by(sex)
*Results ((highly significant)): t =  9.9118  Pr(|T| > |t|) = 0
*men have higher mean 
*(i.e. agree that if a woman earns more it'll certainly cause problems)

* For India
ttest v47 if country_name=="India", by(sex)
*Results ((highly significant)): t =   3.1155 Pr(|T| > |t|) = 0.0018 
*men have higher mean 
*(i.e. agree that if a woman earns more it'll certainly cause problems)


save "WVS_2016.dta", replace


**********
** NES ***
**********


cd "$analysis_datasets"

use "NES_1985_truncated.dta", clear

* Source: NES 1985 Round


* STEP 1. Create Gender dummies *
*--------------------------------

tab q22
rename q22 sex
* Note male = 1, female = 2


* STEP 2. Ttests *
*-----------------

*Political Interest
*-------------------

**Q2. Political interest in 1984

tab q2

tostring q2, gen(Q2)
destring Q2, replace

tab Q2

*Flip ranking order of answer choices->1 = no interest (ease of interpretation)
gen polint = .
replace polint = 1 if Q2 == 3
replace polint = 2 if Q2 == 2
replace polint = 3 if Q2 == 1

tab Q2 polint
*now scores range from 1 "Not interested" to 3 "Very interested" 

ttest Q2, by(sex)
ttest polint, by(sex)
*Results ((highly significant)):  t =   8.5460   Pr(|T| > |t|) = 0



**Q9. How much do you talk about the election

tab q9

tostring q9, gen(Q9)
destring Q9, replace

tab Q9

*Flip ranking order of answer choices
gen poltalk = .
replace poltalk = 1 if Q9 == 3
replace poltalk = 2 if Q9 == 2
replace poltalk = 3 if Q9 == 1

tab Q9 poltalk
*now scores range from 1 "Almost none" to 3 "Talked much" 

tab q9 sex

ttest Q9, by(sex)
ttest poltalk, by(sex)
*Results ((highly significant)):  t =   8.2709  Pr(|T| > |t|) = 0



*Political Participation
*-----------------------

**Q2a. Attend election meetings 

tab q2a

tostring q2a, gen(Q2a)
destring Q2a, replace

tab Q2a


*Flip ranking order of answer choices (for ease of interpretation)
gen elecmt = .
replace elecmt = 1 if Q2a == 3
replace elecmt = 2 if Q2a == 2
replace elecmt = 3 if Q2a == 1

tab Q2a elecmt
*now scores range from 1 "Never" to 3 "Several" 

ttest Q2a, by(sex)
ttest elecmt, by(sex)
*Results ((highly significant)):  t =   9.7693   Pr(|T| > |t|) = 0


**Q9b. Did you work in the campaign 

tab q9b

tostring q9b, gen(Q9b)
destring Q9b, replace

tab Q9b

*Flip ranking order of answer choices (for ease of interpretation)
gen camp = .
replace camp = 1 if Q9b == 3
replace camp = 2 if Q9b == 2
replace camp = 3 if Q9b == 1


tab Q9b camp
*now scores range from 1 "Not at all" to 3 "Much"

tab Q9b sex

ttest Q9b, by(sex)
ttest camp, by(sex)
*Results ((highly significant)):  t =   5.1747   Pr(|T| > |t|) = 0

save "NES_1985.dta", replace


clear


****************
** ISSP 2012 ***
****************


cd "$analysis_datasets"

use "ISSP_2012_truncated.dta", clear

* Source: ISSP 2012 round


* STEP 1. Create Gender dummies *
*--------------------------------

tab SEX

tostring SEX, gen(sex)

destring sex, replace

tab sex

replace sex=. if SEX==9
 
 

* STEP 2. Name Countries *
*-------------------------

sdecode V3, gen(country_name)

tab country_name



* STEP 3. T-tests *
*------------------


**Income Sharing**
*-----------------

**Q2a. Both man and woman should contribute to household income** 

tab V10 

tostring V10, gen(v10)
destring v10, replace 

tab v10

* Treat "Can't choose" and "No answer" as missing values
 
replace v10=. if V10==8
replace v10=. if V10==9

tab v10
*scores vary from 1 "Strongly agree" to 5 "Strongly disagree"  


* Flip ranking order of answer choices (in order to facilitate result interpretation)
 
 replace v10=1 if V10==5
 replace v10=2 if V10==4
 replace v10=4 if V10==2
 replace v10=5 if V10==1

* Include "Spain Only" answers in the full list (i.e. merge ES_V10 in v10) 
 
tab V3 ES_V10
tab ES_V10

tostring ES_V10, gen(es_v10)

tab es_v10

replace v10=1 if ES_V10==4
replace v10=2 if ES_V10==3
replace v10=4 if ES_V10==2
replace v10=5 if ES_V10==1
replace v10=. if v10==0

tab V3 v10
*now scores vary from 1 "Strongly disagree" to 5 "Strongly agree" 

* World
ttest v10, by(sex)
*Results ((highly significant)): t = -17.9108  Pr(|T| > |t|) = 0
*women have significantly higher mean
*(i.e. agree that both man and woman should contribute to household income)

* India
ttest v10 if country_name=="IN-India", by(sex)
*Results ((highly significant)):  t =  -2.6996  Pr(|T| > |t|) = 0.0070

save "ISSP_2012.dta", replace


clear



********************************************************************************
********* MAIN ANALYSIS OF SELF-COLLECTED DATA *********************************
********************************************************************************


cd "$analysis_datasets"

use "MainSample_Meghalaya_final.dta", clear


* CONTROLS

global controls "q61 educ relmaj wealthindex"
global controls_alt "q61 educ relmaj" 
global controlsocpol "q61 educ relmaj wealthindex single cores q65 q31 marriage" 



				***************************
				***************************
				***************************

				
cd "$output_path"
				

		

********************************************************************************
***** MAIN FIGURES & ASSOCIATED APPENDIX TABLES  *******************************
********************************************************************************


		
				
*************************************				
*** FIGURE 3 & APPENDIX TABLE A11 ***				
*************************************				
				
				
* Table 11a
*----------
				
* q38: voting in latest MLA elections

putexcel set A11a_final_2020.xls, replace

*set titles
putexcel C1  = ("Patrilineal") 	E1 = ("Matrilineal") ///
		 A3  = ("Men") 			///
		 A4  = ("Observations") ///
		 A6  = ("Women") 		///
		 A7 = ("Observations")  ///
		 A9 = ("Men - Women")   ///
		 A10 = ("s.e.") 		///
		 A11 = ("p-value")		///
		 A13 = ("Difference (a-b)") ///
		 A14 = ("s.e.") 		///
		 A15 = ("p-value")		
		 
		 
* means
		 
* men
*----

*patrilineal
sum q38_outcome if (q0 == 1 & group == 0)
putexcel C3 = matrix(r(mean)) C4 = matrix(r(N))

*matrilineal
sum q38_outcome if (q0 == 1 & group == 1)
putexcel E3 = matrix(r(mean)) E4 = matrix(r(N))


* women
*------

*patrilineal
sum q38_outcome if (q0 == 2 & group == 0)
putexcel C6 = matrix(r(mean)) C7 = matrix(r(N))

*matrilineal
sum q38_outcome if (q0 == 2 & group == 1)
putexcel E6 = matrix(r(mean)) E7 = matrix(r(N))


* ttests

*patrilineal women vs. men
estpost ttest q38_outcome if (group == 0), by(q0)
putexcel C9 = matrix(e(b)) C10 = matrix(e(se)) C11 = matrix(e(p))

*matrilineal women vs. men
estpost ttest q38_outcome if (group == 1), by(q0)
putexcel E9 = matrix(e(b)) E10 = matrix(e(se)) E11 = matrix(e(p))
				

*FOOTNOTE 9: diff-in-diff test 
* beta = 0.20; p=0.000
eststo: reg q38_outcome i.q0##i.group, robust
local t = _b[2.q0#1.group]/_se[2.q0#1.group]
local p = ttail(e(df_r), `t')
putexcel C13 = matrix(_b[2.q0#1.group]) C14 = (_se[2.q0#1.group]) C15 = (`p')


	
* table 11(b)
*------------

*q34: do you trust legislators to do the right thing for people in Shillong?

putexcel set A11b_final_2020.xls, replace


*set titles
putexcel C1  = ("Patrilineal") 	E1 = ("Matrilineal") ///
		 A3  = ("Men") 			///
		 A4  = ("Observations") ///
		 A6  = ("Women") 		///
		 A7 = ("Observations")  ///
		 A9 = ("Men - Women")   ///
		 A10 = ("s.e.") 		///
		 A11 = ("p-value")		///
		 A13 = ("Difference (a-b)") ///
		 A14 = ("s.e.") 		///
		 A15 = ("p-value")		
		 
* means
		 
* men
*----

*patrilineal
sum q34_outcome if (q0 == 1 & group == 0)
putexcel C3 = matrix(r(mean)) C4 = matrix(r(N))

*matrilineal
sum q34_outcome if (q0 == 1 & group == 1)
putexcel E3 = matrix(r(mean)) E4 = matrix(r(N))


* women
*------

*patrilineal
sum q34_outcome if (q0 == 2 & group == 0)
putexcel C6 = matrix(r(mean)) C7 = matrix(r(N))

*matrilineal
sum q34_outcome if (q0 == 2 & group == 1)
putexcel E6 = matrix(r(mean)) E7 = matrix(r(N))


* ttests

*patrilineal women vs. men
estpost ttest q34_outcome if (group == 0), by(q0)
putexcel C9 = matrix(e(b)) C10 = matrix(e(se)) C11 = matrix(e(p))

*matrilineal women vs. men
estpost ttest q34_outcome if (group == 1), by(q0)
putexcel E9 = matrix(e(b)) E10 = matrix(e(se)) E11 = matrix(e(p))
				
	
*diff-in-diff test for appendix
eststo: reg q34_outcome i.q0##i.group, robust
local t = _b[2.q0#1.group]/_se[2.q0#1.group]
local p = ttail(e(df_r), `t')
putexcel C13 = matrix(_b[2.q0#1.group]) C14 = (_se[2.q0#1.group]) C15 = (`p')



				
** table 11(c)
**------------

*q35:do you trust political parties to do the right thing for people in Shillong?

putexcel set A11c_final_2020.xls, replace

**set titles
putexcel C1  = ("Patrilineal") 	E1 = ("Matrilineal") ///
		 A3  = ("Men") 			///
		 A4  = ("Observations") ///
		 A6  = ("Women") 		///
		 A7 = ("Observations")  ///
		 A9 = ("Men - Women")   ///
		 A10 = ("s.e.") 		///
		 A11 = ("p-value")		///
		 A13 = ("Difference (a-b)") ///
		 A14 = ("s.e.") 		///
		 A15 = ("p-value")		
		 
* means
		 
* men
*----

*patrilineal
sum q35_outcome if (q0 == 1 & group == 0)
putexcel C3 = matrix(r(mean)) C4 = matrix(r(N))

*matrilineal
sum q35_outcome if (q0 == 1 & group == 1)
putexcel E3 = matrix(r(mean)) E4 = matrix(r(N))


* women
*------

*patrilineal
sum q35_outcome if (q0 == 2 & group == 0)
putexcel C6 = matrix(r(mean)) C7 = matrix(r(N))

*matrilineal
sum q35_outcome if (q0 == 2 & group == 1)
putexcel E6 = matrix(r(mean)) E7 = matrix(r(N))


* ttests

*patrilineal women vs. men
estpost ttest q35_outcome if (group == 0), by(q0)
putexcel C9 = matrix(e(b)) C10 = matrix(e(se)) C11 = matrix(e(p))

*matrilineal women vs. men
estpost ttest q35_outcome if (group == 1), by(q0)
putexcel E9 = matrix(e(b)) E10 = matrix(e(se)) E11 = matrix(e(p))
	
	

*diff-in-diff test for appendix
eststo: reg q35_outcome i.q0##i.group, robust

local t = _b[2.q0#1.group]/_se[2.q0#1.group]
local p = ttail(e(df_r), `t')
putexcel C13 = matrix(_b[2.q0#1.group]) C14 = (_se[2.q0#1.group]) C15 = (`p')

	

*table 11(d)
*-----------

*q28: possible to hold local politicians accountable?

putexcel set A11d_final_2020.xls, replace

*set titles
putexcel C1  = ("Patrilineal") 	E1 = ("Matrilineal") ///
		 A3  = ("Men") 			///
		 A4  = ("Observations") ///
		 A6  = ("Women") 		///
		 A7 = ("Observations")  ///
		 A9 = ("Men - Women")   ///
		 A10 = ("s.e.") 		///
		 A11 = ("p-value")		///
		 A13 = ("Difference (a-b)") ///
		 A14 = ("s.e.") 		///
		 A15 = ("p-value")		
		 
* means
		 
* men
*----

*patrilineal
sum q28_outcome if (q0 == 1 & group == 0)
putexcel C3 = matrix(r(mean)) C4 = matrix(r(N))

*matrilineal
sum q28_outcome if (q0 == 1 & group == 1)
putexcel E3 = matrix(r(mean)) E4 = matrix(r(N))


* women
*------

*patrilineal
sum q28_outcome if (q0 == 2 & group == 0)
putexcel C6 = matrix(r(mean)) C7 = matrix(r(N))

*matrilineal
sum q28_outcome if (q0 == 2 & group == 1)
putexcel E6 = matrix(r(mean)) E7 = matrix(r(N))


* ttests

*patrilineal women vs. men
estpost ttest q28_outcome if (group == 0), by(q0)
putexcel C9 = matrix(e(b)) C10 = matrix(e(se)) C11 = matrix(e(p))

*matrilineal women vs. men
estpost ttest q28_outcome if (group == 1), by(q0)
putexcel E9 = matrix(e(b)) E10 = matrix(e(se)) E11 = matrix(e(p))
				
	
*FOOTNOTE 10:  diff-in-diff test 
eststo: reg q28_outcome i.q0##i.group, robust
local t = _b[2.q0#1.group]/_se[2.q0#1.group]
local p = ttail(e(df_r), `t')
putexcel C13 = matrix(_b[2.q0#1.group]) C14 = (_se[2.q0#1.group]) C15 = (`p')


	
	

				***************************
				***************************
				***************************		
				
				
				

********************************************************************************
***** MAIN TABLES **************************************************************
********************************************************************************

				

* Table 1: Personal cost treatment (q8)
*--------------------------------------

eststo clear

*pat men

eststo: reg q8_outcome q8_treat $controls ///
	if (q0 == 1 & group == 0), robust

	
*pat women

eststo: reg q8_outcome q8_treat $controls ///
	if (q0 == 2 & group == 0), robust
	
	
*mat men

eststo: reg q8_outcome q8_treat $controls ///
	if (q0 == 1 & group == 1), robust

	
*mat women

eststo: reg q8_outcome q8_treat $controls ///
	if (q0 == 2 & group == 1), robust
	
	
esttab using "T1_final_2020.csv", b(2) se(2) replace ///
		label title(Effect of Personal Cost Treatment on Policy Preferences) ///
		mtitle("Pat Men" "Pat Women" "Mat Men" "Mat Women") ///
		keep(q8_treat _cons) order(q8_treat _cons) gaps ///
		varlabels(q8_treat "Explicit cost to policy" _cons ///
		"Constant (control mean)") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear		


* Comparison of the gender gap across cultures, footnote 13, pg. 18

ologit q8_outcome i.q8_treat##i.q0##i.group $controls, robust
*q8_treat#q0#group = -1.976694, se = .9378966, p-val = 0.035

			

				***************************
				***************************
				***************************		
				
	
	

* Table 2: Main Anaysis, postcard treatment (q77)
*------------------------------------------------

* NOTE: Our outcome here is policy preferences expressed by postcard senders.

eststo clear

*pat men
eststo: reg q77_outcome q77_treat $controls ///
	if (q0 == 1 & group == 0), robust

*pat women
eststo: reg q77_outcome q77_treat $controls ///
	if (q0 == 2 & group == 0), robust
	
*mat men
eststo: reg q77_outcome q77_treat $controls ///
	if (q0 == 1 & group == 1), robust

*mat women
eststo: reg q77_outcome q77_treat $controls ///
	if (q0 == 2 & group == 1), robust
	
	
esttab using "T2_final_2020.csv", b(2) se(2) replace ///
		label title(Effect of Postcard Treatment on Policy Preferences) ///
		mtitle("Pat Men" "Pat Women" "Mat Men" "Mat Women") ///
		keep(q77_treat _cons) order(q77_treat _cons) gaps ///
		varlabels(q77_treat "Explicit cost to policy" _cons ///
		"Constant (control mean)") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear		
	
	
* Compare gender gap across cultures, footnote 	18, [pg. 21]
ologit q77_outcome i.q77_treat##i.q0##i.group $controls, robust
*q77_treat#q0#group = -3.006387, se =  1.591035 , p-val = 0.059



				***************************
				***************************
				***************************	
								
		


* Table 3: Preferences about decision making (q19), main analysis with controls
*------------------------------------------------------------------------------

eststo clear

*pat men
eststo: reg q19_outcome q19_treat $controls if (q0 == 1 & group == 0 & ///
		(q19_treat == 1 | q19_treat == 3)), robust
			
*pat women
eststo: reg q19_outcome q19_treat $controls if (q0 == 2 & group == 0 & ///
		(q19_treat == 1 | q19_treat == 3)), robust

*mat men
eststo: reg q19_outcome q19_treat $controls if (q0 == 1 & group == 1 & ///
		(q19_treat == 1 | q19_treat == 3)), robust

*mat women
eststo: reg q19_outcome q19_treat $controls if (q0 == 2 & group == 1 & ///
		(q19_treat == 1 | q19_treat == 3)), robust
		
esttab using "T3_final_2020.csv", b(2) se(2) replace ///
		label title(Intra-HH decision-making Preferences ///
		Conditional on Income Contribution; baseline = both partners earn) ///
		mtitle("Pat Men" "Pat Women" "Mat Men" "Mat Women") ///
		keep(q19_treat _cons) ///
		order(q19_treat _cons) gaps ///
		varlabels(q19_treat "Women earn more" _cons ///
		"Constant (control mean)") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear		
	
	

				***************************
				***************************
				***************************	
				
				

* Table 4: public goods distribution (q10)
*-----------------------------------------

* Compare baseline with treatment where head gets more money 

eststo clear


*pat men
eststo: reg q10_outcome q10_treat $controls ///
	if (q0 == 1 & group == 0 & ///
	(q10_treat == 0 | q10_treat == 1)), robust

	
*pat women
eststo: reg q10_outcome q10_treat $controls ///
	if (q0 == 2 & group == 0  & ///
	(q10_treat == 0 | q10_treat == 1)), robust
	
	
*mat men
eststo: reg q10_outcome q10_treat $controls ///
	if (q0 == 1 & group == 1 & ///
	(q10_treat == 0 | q10_treat == 1)), robust

	
*mat women
eststo: reg q10_outcome q10_treat $controls ///
	if (q0 == 2 & group == 1 & ///
	(q10_treat == 0 | q10_treat == 1)), robust

	
esttab using "T4_final_2020.csv", b(2) se(2) replace ///
		label title(Effect of Costly Gov't Distribution on Policy Preferences) ///
		mtitle("Pat Men" "Pat Women" "Mat Men" "Mat Women") ///
		keep(q10_treat _cons) order(q10_treat _cons) gaps ///
		varlabels(q10_treat "Costly gov't dist" _cons ///
		"Constant (control mean)") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear		
	
			
				***************************
				***************************
				***************************	
				
						
				


********************************************************************************
***** APPENDIX TABLES **********************************************************
********************************************************************************

				
	

* Appendix A6: Summary Statistics by Gender *
*--------------------------------------------


*create table
putexcel set A6_final_2020.xls, replace

*set titles
putexcel C1  = ("Matrilineal") 	F1 = ("Patrilineal")  I1 = ("Difference") ///
		 A2  = ("Variable") C2 = ("mean") D2 = ("N") F2 = ("mean") G2 = ("N") ///
		 I2  = ("mean") 		J2 =("SE") 	  	K2 = ("p-value") ///
		 A3  = ("Men") ///
		 A4  = ("Number siblings") 			A5  = ("Age") ///
		 A6  = ("Marriage year") 			A7  = ("Number daughters") ///
		 A8  = ("Any education") ///
		 A9 = ("Prim education") 		A10 = ("High wages (2 if > average)") ///
		 A11 = ("Culturally-determined variables") ///
		 A12 = ("Wealth index") ///
		 A13 = ("Land (title)") ///			
		 A15 = ("Women") ///
		 A16 = ("Number siblings") 			A17 = ("Age") ///
		 A18 = ("Marriage year") 			A19 = ("Number daughters") ///
		 A20 = ("Any education") ///
		 A21 = ("Prim education") 			A22 = ("High wages (2 if > average)") ///
		 A23 = ("Culturally-determined variables") ///
		 A24 = ("Wealth index") ///
		 A25 = ("Land (title)") 		
		 
* Part 1. Means
*--------------		 

* men
*----

* matrilineal

*number siblings
sum q51 		if (q0 == 1 & group == 1)
putexcel C4 = matrix(r(mean)) D4 = matrix(r(N))  

*age
sum q61 		if (q0 == 1 & group == 1)
putexcel C5 = matrix(r(mean)) D5 = matrix(r(N)) 

*marriage year
sum q69 		if (q0 == 1 & group == 1)
putexcel C6 = matrix(r(mean)) D6 = matrix(r(N))  

*number daughters
sum q73 		if (q0 == 1 & group == 1)
putexcel C7 = matrix(r(mean)) D7 = matrix(r(N))  

*education
sum educany 	if (q0 == 1 & group == 1)
putexcel C8 = matrix(r(mean)) D8 = matrix(r(N))  

*primary education
sum educ		if (q0 == 1 & group == 1)
putexcel C9 = matrix(r(mean)) D9 = matrix(r(N))

*wages, all
sum wages 		if (q0 == 1 & group == 1)
putexcel C10 = matrix(r(mean)) D10 = matrix(r(N))    
 
*wealth index
sum wealthindex if (q0 == 1 & group == 1)
putexcel C12 = matrix(r(mean)) D12 = matrix(r(N))  

*land title
sum landt 		if (q0 == 1 & group == 1)
putexcel C13 = matrix(r(mean)) D13 = matrix(r(N))   




* patrilineal

*number siblings
sum q51 		if (q0 == 1 & group == 0)
putexcel F4 = matrix(r(mean)) G4 = matrix(r(N))  

*age
sum q61 		if (q0 == 1 & group == 0)
putexcel F5 = matrix(r(mean)) G5 = matrix(r(N)) 

*marriage year
sum q69 		if (q0 == 1 & group == 0)
putexcel F6 = matrix(r(mean)) G6 = matrix(r(N))  

*number daughters
sum q73 		if (q0 == 1 & group == 0)
putexcel F7 = matrix(r(mean)) G7 = matrix(r(N))  

*education
sum educany 	if (q0 == 1 & group == 0)
putexcel F8 = matrix(r(mean)) G8 = matrix(r(N))  

*education beyond primary
sum educ		if (q0 == 1 & group == 0)
putexcel F9 = matrix(r(mean)) G9 = matrix(r(N))  

*wages, all
sum wages 		if (q0 == 1 & group == 0)
putexcel F10 = matrix(r(mean)) G10 = matrix(r(N)) 

*wealth index
sum wealthindex if (q0 == 1 & group == 0)
putexcel F12 = matrix(r(mean)) G12 = matrix(r(N))   

*land title
sum landt 		if (q0 == 1 & group == 0)
putexcel F13 = matrix(r(mean)) G13 = matrix(r(N))   

   



* women
*------
* matrilineal

*number siblings
sum q51 		if (q0 == 2 & group == 1)
putexcel C16 = matrix(r(mean)) D16 = matrix(r(N))  

*age
sum q61 		if (q0 == 2 & group == 1)
putexcel C17 = matrix(r(mean)) D17 = matrix(r(N)) 

*marriage year
sum q69 		if (q0 == 2 & group == 1)
putexcel C18 = matrix(r(mean)) D18 = matrix(r(N))  

*number daughters
sum q73 		if (q0 == 2 & group == 1)
putexcel C19 = matrix(r(mean)) D19 = matrix(r(N))  

*education
sum educany 	if (q0 == 2 & group == 1)
putexcel C20 = matrix(r(mean)) D20 = matrix(r(N))  

sum educ		if (q0 == 2 & group == 1)
putexcel C21 = matrix(r(mean)) D21 = matrix(r(N)) 

*wages, all
sum wages 		if (q0 == 2 & group == 1)
putexcel C22 = matrix(r(mean)) D22 = matrix(r(N))  

*wealth index
sum wealthindex if (q0 == 2 & group == 1)
putexcel C24 = matrix(r(mean)) D24 = matrix(r(N))  

*land title
sum landt 		if (q0 == 2 & group == 1)
putexcel C25 = matrix(r(mean)) D25 = matrix(r(N))   

  



* patrilineal

*number siblings
sum q51 		if (q0 == 2 & group == 0)
putexcel F16 = matrix(r(mean)) G16 = matrix(r(N))  

*age
sum q61 		if (q0 == 2 & group == 0)
putexcel F17 = matrix(r(mean)) G17 = matrix(r(N)) 

*marriage year
sum q69 		if (q0 == 2 & group == 0)
putexcel F18 = matrix(r(mean)) G18 = matrix(r(N))  

*number daughters
sum q73 		if (q0 == 2 & group == 0)
putexcel F19 = matrix(r(mean)) G19 = matrix(r(N))  

*education
sum educany 	if (q0 == 2 & group == 0)
putexcel F20 = matrix(r(mean)) G20 = matrix(r(N))  

sum educ		if (q0 == 2 & group == 0)
putexcel F21 = matrix(r(mean)) G21 = matrix(r(N))  

*wages, all
sum wages 		if (q0 == 2 & group == 0)
putexcel F22 = matrix(r(mean)) G22 = matrix(r(N)) 

*wealth index
sum wealthindex if (q0 == 2 & group == 0)
putexcel F24 = matrix(r(mean)) G24 = matrix(r(N)) 

*land title
sum landt 		 if (q0 == 2 & group == 0)
putexcel F25 = matrix(r(mean)) G25 = matrix(r(N))   

  



* Part 2. Ttests: differences in means
*-------------------------------------	 
		 
* men
*----

*number siblings
estpost ttest q51 if q0 == 1, by (group)
putexcel I4 = matrix(e(b)) J4 = matrix(e(se)) K4 = matrix(e(p)) 

*age
estpost ttest q61 if q0 == 1, by (group)
putexcel I5 = matrix(e(b)) J5 = matrix(e(se)) K5 = matrix(e(p)) 

*marriage year
estpost ttest q69 if q0 == 1, by (group)
putexcel I6 = matrix(e(b)) J6 = matrix(e(se)) K6 = matrix(e(p)) 

*number daughters
estpost ttest q73 if q0 == 1, by (group)
putexcel I7 = matrix(e(b)) J7 = matrix(e(se)) K7 = matrix(e(p)) 

*education
estpost ttest educany 	if q0 == 1, by(group)
putexcel I8 = matrix(e(b)) J8 = matrix(e(se)) K8 = matrix(e(p)) 

estpost ttest educ		if q0 == 1, by (group)
putexcel I9 = matrix(e(b)) J9 = matrix(e(se)) K9 = matrix(e(p)) 

*wages
estpost ttest wages if q0 == 1, by (group)
putexcel I10 = matrix(e(b)) J10 = matrix(e(se)) K10 = matrix(e(p)) 

*wealth index
estpost ttest wealthindex if q0 == 1, by (group)
putexcel I12 = matrix(e(b)) J12 = matrix(e(se)) K12 = matrix(e(p))  

*land title
estpost ttest landt if q0 == 1, by (group)
putexcel I13 = matrix(e(b)) J13 = matrix(e(se)) K13 = matrix(e(p)) 

 


* women
*------

*number siblings
estpost ttest q51 if q0 == 2, by (group)
putexcel I16 = matrix(e(b)) J16 = matrix(e(se)) K16 = matrix(e(p)) 

*age
estpost ttest q61 if q0 == 2, by (group)
putexcel I17 = matrix(e(b)) J17 = matrix(e(se)) K17 = matrix(e(p)) 

*marriage year
estpost ttest q69 if q0 == 2, by (group)
putexcel I18 = matrix(e(b)) J18 = matrix(e(se)) K18 = matrix(e(p)) 

*number daughters
estpost ttest q73 if q0 == 2, by (group)
putexcel I19 = matrix(e(b)) J19 = matrix(e(se)) K19 = matrix(e(p)) 

*education
estpost ttest educany 	if q0 == 2, by(group)
putexcel I20 = matrix(e(b)) J20 = matrix(e(se)) K20 = matrix(e(p)) 

*primary education
estpost ttest educ		if q0 == 2, by (group)
putexcel I21 = matrix(e(b)) J21 = matrix(e(se)) K21 = matrix(e(p)) 

*wages
estpost ttest wages if q0 == 2, by (group)
putexcel I22 = matrix(e(b)) J22 = matrix(e(se)) K22 = matrix(e(p)) 

*wealth index
estpost ttest wealthindex if q0 == 2, by (group)
putexcel I24 = matrix(e(b)) J24 = matrix(e(se)) K24 = matrix(e(p)) 

*land title
estpost ttest landt if q0 == 2, by (group)
putexcel I25 = matrix(e(b)) J25 = matrix(e(se)) K25 = matrix(e(p)) 





				***************************
				***************************
				***************************				
				
				
								
				
				
* Appendix table A7: personal cost treatment (q8), no controls
*-------------------------------------------------------------

* Balance tests
*--------------

*create table
putexcel set A7_final_2020.xls, replace

*set titles
putexcel C1  = ("Control") 	F1 = ("Treatment")  I1 = ("Difference") ///
		 A2  = ("Variable") C2 = ("mean") D2 = ("N") F2 = ("mean") G2 = ("N") ///
		 I2  = ("mean") 		J2 =("SE") 	  	K2 = ("p-value") ///
		 A3  = ("Men") ///
		 A4  = ("Number siblings") 			A5  = ("Age") ///
		 A6  = ("Marriage year") 			A7  = ("Number daughters") ///
		 A8  = ("Any education") ///
		 A9 = ("Prim education") 		A10 = ("High wages (2 if > average)") ///
		 A11 = ("Culturally-determined variables") ///
		 A12 = ("Wealth index") ///
		 A13 = ("Land (title)") ///	
		
		 
* Part 1. Means
*--------------		 

* control

*number siblings
sum q51 		if (q8_treat == 0)
putexcel C4 = matrix(r(mean)) D4 = matrix(r(N))  

*age
sum q61 		if (q8_treat == 0)
putexcel C5 = matrix(r(mean)) D5 = matrix(r(N)) 

*marriage year
sum q69 		if (q8_treat == 0)
putexcel C6 = matrix(r(mean)) D6 = matrix(r(N))  

*number daughters
sum q73 		if (q8_treat == 0)
putexcel C7 = matrix(r(mean)) D7 = matrix(r(N))  

*education
sum educany 	if (q8_treat == 0)
putexcel C8 = matrix(r(mean)) D8 = matrix(r(N))  

*primary education
sum educ		if (q8_treat == 0)
putexcel C9 = matrix(r(mean)) D9 = matrix(r(N))

*wages, all
sum wages 		if (q8_treat == 0)
putexcel C10 = matrix(r(mean)) D10 = matrix(r(N))    
 
*wealth index
sum wealthindex if (q8_treat == 0)
putexcel C12 = matrix(r(mean)) D12 = matrix(r(N))  

*land title
sum landt 		if (q8_treat == 0)
putexcel C13 = matrix(r(mean)) D13 = matrix(r(N))   


* treatment

*number siblings
sum q51 		if (q8_treat == 1)
putexcel F4 = matrix(r(mean)) G4 = matrix(r(N))  

*age
sum q61 		if (q8_treat == 1)
putexcel F5 = matrix(r(mean)) G5 = matrix(r(N)) 

*marriage year
sum q69 		if (q8_treat == 1)
putexcel F6 = matrix(r(mean)) G6 = matrix(r(N))  

*number daughters
sum q73 		if (q8_treat == 1)
putexcel F7 = matrix(r(mean)) G7 = matrix(r(N))  

*education
sum educany 	if (q8_treat == 1)
putexcel F8 = matrix(r(mean)) G8 = matrix(r(N))  

*education beyond primary
sum educ		if (q8_treat == 1)
putexcel F9 = matrix(r(mean)) G9 = matrix(r(N))  

*wages, all
sum wages 		if (q8_treat == 1)
putexcel F10 = matrix(r(mean)) G10 = matrix(r(N)) 

*wealth index
sum wealthindex if (q8_treat == 1)
putexcel F12 = matrix(r(mean)) G12 = matrix(r(N))   

*land title
sum landt 		if (q8_treat == 1)
putexcel F13 = matrix(r(mean)) G13 = matrix(r(N))   
  
   

* Part 2. T-tests: differences in means
*--------------------------------------	 

*number siblings
estpost ttest q51, by (q8_treat) uneq
putexcel I4 = matrix(e(b)) J4 = matrix(e(se)) K4 = matrix(e(p)) 

*age
estpost ttest q61, by (q8_treat) uneq
putexcel I5 = matrix(e(b)) J5 = matrix(e(se)) K5 = matrix(e(p)) 

*marriage year
estpost ttest q69, by (q8_treat) uneq
putexcel I6 = matrix(e(b)) J6 = matrix(e(se)) K6 = matrix(e(p)) 

*number daughters
estpost ttest q73, by (q8_treat) uneq
putexcel I7 = matrix(e(b)) J7 = matrix(e(se)) K7 = matrix(e(p)) 

*education
estpost ttest educany, by (q8_treat) uneq
putexcel I8 = matrix(e(b)) J8 = matrix(e(se)) K8 = matrix(e(p)) 

estpost ttest educ,		by (q8_treat) uneq
putexcel I9 = matrix(e(b)) J9 = matrix(e(se)) K9 = matrix(e(p)) 

*wages
estpost ttest wages, by (q8_treat) uneq
putexcel I10 = matrix(e(b)) J10 = matrix(e(se)) K10 = matrix(e(p)) 

*wealth index
estpost ttest wealthindex, by (q8_treat) uneq
putexcel I12 = matrix(e(b)) J12 = matrix(e(se)) K12 = matrix(e(p))  

*land title
estpost ttest landt, by (q8_treat) uneq
putexcel I13 = matrix(e(b)) J13 = matrix(e(se)) K13 = matrix(e(p)) 

  


				***************************
				***************************
				***************************						
				

				
* Appendix table A8: Balance test, Randomization for Table 2 treatment
*---------------------------------------------------------------------

*create table
putexcel set A8_final_2020.xls, replace

*set titles
putexcel C1  = ("Control") 	F1 = ("Treatment")  I1 = ("Difference") ///
		 A2  = ("Variable") C2 = ("mean") D2 = ("N") F2 = ("mean") G2 = ("N") ///
		 I2  = ("mean") 		J2 =("SE") 	  	K2 = ("p-value") ///
		 A3  = ("Men") ///
		 A4  = ("Number siblings") 			A5  = ("Age") ///
		 A6  = ("Marriage year") 			A7  = ("Number daughters") ///
		 A8  = ("Any education") ///
		 A9 = ("Prim education") 		A10 = ("High wages (2 if > average)") ///
		 A11 = ("Culturally-determined variables") ///
		 A12 = ("Wealth index") ///
		 A13 = ("Land (title)")
		
		 
* Part 1. Means
*--------------		 

* control

*number siblings
sum q51 		if (q77_treat == 0)
putexcel C4 = matrix(r(mean)) D4 = matrix(r(N))  

*age
sum q61 		if (q77_treat == 0)
putexcel C5 = matrix(r(mean)) D5 = matrix(r(N)) 

*marriage year
sum q69 		if (q77_treat == 0)
putexcel C6 = matrix(r(mean)) D6 = matrix(r(N))  

*number daughters
sum q73 		if (q77_treat == 0)
putexcel C7 = matrix(r(mean)) D7 = matrix(r(N))  

*education
sum educany 	if (q77_treat == 0)
putexcel C8 = matrix(r(mean)) D8 = matrix(r(N))  

*primary education
sum educ		if (q77_treat == 0)
putexcel C9 = matrix(r(mean)) D9 = matrix(r(N))

*wages, all
sum wages 		if (q77_treat == 0)
putexcel C10 = matrix(r(mean)) D10 = matrix(r(N))    
 
*wealth index
sum wealthindex if (q77_treat == 0)
putexcel C12 = matrix(r(mean)) D12 = matrix(r(N))  

*land title
sum landt 		if (q77_treat == 0)
putexcel C13 = matrix(r(mean)) D13 = matrix(r(N))   


* treatment

*number siblings
sum q51 		if (q77_treat == 1)
putexcel F4 = matrix(r(mean)) G4 = matrix(r(N))  

*age
sum q61 		if (q77_treat == 1)
putexcel F5 = matrix(r(mean)) G5 = matrix(r(N)) 

*marriage year
sum q69 		if (q77_treat == 1)
putexcel F6 = matrix(r(mean)) G6 = matrix(r(N))  

*number daughters
sum q73 		if (q77_treat == 1)
putexcel F7 = matrix(r(mean)) G7 = matrix(r(N))  

*education
sum educany 	if (q77_treat == 1)
putexcel F8 = matrix(r(mean)) G8 = matrix(r(N))  

*education beyond primary
sum educ		if (q77_treat == 1)
putexcel F9 = matrix(r(mean)) G9 = matrix(r(N))  

*wages, all
sum wages 		if (q77_treat == 1)
putexcel F10 = matrix(r(mean)) G10 = matrix(r(N)) 

*wealth index
sum wealthindex if (q77_treat == 1)
putexcel F12 = matrix(r(mean)) G12 = matrix(r(N))   

*land title
sum landt 		if (q77_treat == 1)
putexcel F13 = matrix(r(mean)) G13 = matrix(r(N))   
 

* Part 2. Ttests: differences in means
*-------------------------------------	 

*number siblings
estpost ttest q51, by (q77_treat) uneq
putexcel I4 = matrix(e(b)) J4 = matrix(e(se)) K4 = matrix(e(p)) 

*age
estpost ttest q61, by (q77_treat) uneq
putexcel I5 = matrix(e(b)) J5 = matrix(e(se)) K5 = matrix(e(p)) 

*marriage year
estpost ttest q69, by (q77_treat) uneq
putexcel I6 = matrix(e(b)) J6 = matrix(e(se)) K6 = matrix(e(p)) 

*number daughters
estpost ttest q73, by (q77_treat) uneq
putexcel I7 = matrix(e(b)) J7 = matrix(e(se)) K7 = matrix(e(p)) 

*education
estpost ttest educany, by (q77_treat) uneq
putexcel I8 = matrix(e(b)) J8 = matrix(e(se)) K8 = matrix(e(p)) 

estpost ttest educ,		by (q77_treat) uneq
putexcel I9 = matrix(e(b)) J9 = matrix(e(se)) K9 = matrix(e(p)) 

*wages
estpost ttest wages, by (q77_treat) uneq
putexcel I10 = matrix(e(b)) J10 = matrix(e(se)) K10 = matrix(e(p)) 

*wealth index
estpost ttest wealthindex, by (q77_treat) uneq
putexcel I12 = matrix(e(b)) J12 = matrix(e(se)) K12 = matrix(e(p))  

*land title
estpost ttest landt, by (q77_treat) uneq
putexcel I13 = matrix(e(b)) J13 = matrix(e(se)) K13 = matrix(e(p)) 




				***************************
				***************************
				***************************				
		

* Appendix A9: Balance tests for Table 3 treatment randomization
*---------------------------------------------------------------	

*create table
putexcel set A9_final_2020.xls, replace

*set titles
putexcel C1  = ("Control") 	F1 = ("Treatment")  I1 = ("Difference") ///
		 A2  = ("Variable") C2 = ("mean") D2 = ("N") F2 = ("mean") G2 = ("N") ///
		 I2  = ("mean") 		J2 =("SE") 	  	K2 = ("p-value") ///
		 A3  = ("Men") ///
		 A4  = ("Number siblings") 			A5  = ("Age") ///
		 A6  = ("Marriage year") 			A7  = ("Number daughters") ///
		 A8  = ("Any education") ///
		 A9 = ("Prim education") 		A10 = ("High wages (2 if > average)") ///
		 A11 = ("Culturally-determined variables") ///
		 A12 = ("Wealth index") ///
		 A13 = ("Land (title)") 
		
		 
* Part 1. Means
*--------------		 

* control

*number siblings
sum q51 		if (q19_treat == 1)
putexcel C4 = matrix(r(mean)) D4 = matrix(r(N))  

*age
sum q61 		if (q19_treat == 1)
putexcel C5 = matrix(r(mean)) D5 = matrix(r(N)) 

*marriage year
sum q69 		if (q19_treat == 1)
putexcel C6 = matrix(r(mean)) D6 = matrix(r(N))  

*number daughters
sum q73 		if (q19_treat == 1)
putexcel C7 = matrix(r(mean)) D7 = matrix(r(N))  

*education
sum educany 	if (q19_treat == 1)
putexcel C8 = matrix(r(mean)) D8 = matrix(r(N))  

*primary education
sum educ		if (q19_treat == 1)
putexcel C9 = matrix(r(mean)) D9 = matrix(r(N))

*wages, all
sum wages 		if (q19_treat == 1)
putexcel C10 = matrix(r(mean)) D10 = matrix(r(N))    
 
*wealth index
sum wealthindex if (q19_treat == 1)
putexcel C12 = matrix(r(mean)) D12 = matrix(r(N))  

*land title
sum landt 		if (q19_treat == 1)
putexcel C13 = matrix(r(mean)) D13 = matrix(r(N))   


* treatment

*number siblings
sum q51 		if (q19_treat == 3)
putexcel F4 = matrix(r(mean)) G4 = matrix(r(N))  

*age
sum q61 		if (q19_treat == 3)
putexcel F5 = matrix(r(mean)) G5 = matrix(r(N)) 

*marriage year
sum q69 		if (q19_treat == 3)
putexcel F6 = matrix(r(mean)) G6 = matrix(r(N))  

*number daughters
sum q73 		if (q19_treat == 3)
putexcel F7 = matrix(r(mean)) G7 = matrix(r(N))  

*education
sum educany 	if (q19_treat == 3)
putexcel F8 = matrix(r(mean)) G8 = matrix(r(N))  

*education beyond primary
sum educ		if (q19_treat == 3)
putexcel F9 = matrix(r(mean)) G9 = matrix(r(N))  

*wages, all
sum wages 		if (q19_treat == 3)
putexcel F10 = matrix(r(mean)) G10 = matrix(r(N)) 

*wealth index
sum wealthindex if (q19_treat == 3)
putexcel F12 = matrix(r(mean)) G12 = matrix(r(N))   

*land title
sum landt 		if (q19_treat == 3)
putexcel F13 = matrix(r(mean)) G13 = matrix(r(N))   
   

* Part 2. Ttests: differences in means
*-------------------------------------	 


*number siblings
estpost ttest q51 if (q19_treat == 1 | q19_treat == 3), by (q19_treat) uneq
putexcel I4 = matrix(e(b)) J4 = matrix(e(se)) K4 = matrix(e(p)) 

*age
estpost ttest q61 if (q19_treat == 1 | q19_treat == 3), by (q19_treat) uneq
putexcel I5 = matrix(e(b)) J5 = matrix(e(se)) K5 = matrix(e(p)) 

*marriage year
estpost ttest q69 if (q19_treat == 1 | q19_treat == 3), by (q19_treat) uneq
putexcel I6 = matrix(e(b)) J6 = matrix(e(se)) K6 = matrix(e(p)) 

*number daughters
estpost ttest q73 if (q19_treat == 1 | q19_treat == 3), by (q19_treat) uneq
putexcel I7 = matrix(e(b)) J7 = matrix(e(se)) K7 = matrix(e(p)) 

*education
estpost ttest educany if (q19_treat == 1 | q19_treat == 3), by (q19_treat) uneq
putexcel I8 = matrix(e(b)) J8 = matrix(e(se)) K8 = matrix(e(p)) 

estpost ttest educ if (q19_treat == 1 | q19_treat == 3),		by (q19_treat) uneq
putexcel I9 = matrix(e(b)) J9 = matrix(e(se)) K9 = matrix(e(p)) 

*wages
estpost ttest wages if (q19_treat == 1 | q19_treat == 3), by (q19_treat) uneq
putexcel I10 = matrix(e(b)) J10 = matrix(e(se)) K10 = matrix(e(p)) 

*wealth index
estpost ttest wealthindex if (q19_treat == 1 | q19_treat == 3), by (q19_treat) uneq
putexcel I12 = matrix(e(b)) J12 = matrix(e(se)) K12 = matrix(e(p))  

*land title
estpost ttest landt if (q19_treat == 1 | q19_treat == 3), by (q19_treat) uneq
putexcel I13 = matrix(e(b)) J13 = matrix(e(se)) K13 = matrix(e(p)) 


	
		
				***************************
				***************************
				***************************	
				
				

			
			
* Appendix Table A10: Balance tests for Table 4 treatment randomization
*----------------------------------------------------------------------	

*create table
putexcel set A10_final_2020.xls, replace

*set titles
putexcel C1  = ("Control") 	F1 = ("Treatment")  I1 = ("Difference") ///
		 A2  = ("Variable") C2 = ("mean") D2 = ("N") F2 = ("mean") G2 = ("N") ///
		 I2  = ("mean") 		J2 =("SE") 	  	K2 = ("p-value") ///
		 A3  = ("Men") ///
		 A4  = ("Number siblings") 			A5  = ("Age") ///
		 A6  = ("Marriage year") 			A7  = ("Number daughters") ///
		 A8  = ("Any education") ///
		 A9 = ("Prim education") 		A10 = ("High wages (2 if > average)") ///
		 A11 = ("Culturally-determined variables") ///
		 A12 = ("Wealth index") ///
		 A13 = ("Land (title)")
		
		 
* Part 1. Means
*--------------		 

* control

*number siblings
sum q51 		if (q10_treat == 0)
putexcel C4 = matrix(r(mean)) D4 = matrix(r(N))  

*age
sum q61 		if (q10_treat == 0)
putexcel C5 = matrix(r(mean)) D5 = matrix(r(N)) 

*marriage year
sum q69 		if (q10_treat == 0)
putexcel C6 = matrix(r(mean)) D6 = matrix(r(N))  

*number daughters
sum q73 		if (q10_treat == 0)
putexcel C7 = matrix(r(mean)) D7 = matrix(r(N))  

*education
sum educany 	if (q10_treat == 0)
putexcel C8 = matrix(r(mean)) D8 = matrix(r(N))  

*primary education
sum educ		if (q10_treat == 0)
putexcel C9 = matrix(r(mean)) D9 = matrix(r(N))

*wages, all
sum wages 		if (q10_treat == 0)
putexcel C10 = matrix(r(mean)) D10 = matrix(r(N))    
 
*wealth index
sum wealthindex if (q10_treat == 0)
putexcel C12 = matrix(r(mean)) D12 = matrix(r(N))  

*land title
sum landt 		if (q10_treat == 0)
putexcel C13 = matrix(r(mean)) D13 = matrix(r(N))   


* treatment

*number siblings
sum q51 		if (q10_treat == 1)
putexcel F4 = matrix(r(mean)) G4 = matrix(r(N))  

*age
sum q61 		if (q10_treat == 1)
putexcel F5 = matrix(r(mean)) G5 = matrix(r(N)) 

*marriage year
sum q69 		if (q10_treat == 1)
putexcel F6 = matrix(r(mean)) G6 = matrix(r(N))  

*number daughters
sum q73 		if (q10_treat == 1)
putexcel F7 = matrix(r(mean)) G7 = matrix(r(N))  

*education
sum educany 	if (q10_treat == 1)
putexcel F8 = matrix(r(mean)) G8 = matrix(r(N))  

*education beyond primary
sum educ		if (q10_treat == 1)
putexcel F9 = matrix(r(mean)) G9 = matrix(r(N))  

*wages, all
sum wages 		if (q10_treat == 1)
putexcel F10 = matrix(r(mean)) G10 = matrix(r(N)) 

*wealth index
sum wealthindex if (q10_treat == 1)
putexcel F12 = matrix(r(mean)) G12 = matrix(r(N))   

*land title
sum landt 		if (q10_treat == 1)
putexcel F13 = matrix(r(mean)) G13 = matrix(r(N))   
 

* Part 2. Ttests: differences in means
*-------------------------------------	 


*number siblings
estpost ttest q51 if (q10_treat == 0 | q10_treat == 1), by (q10_treat) uneq
putexcel I4 = matrix(e(b)) J4 = matrix(e(se)) K4 = matrix(e(p)) 

*age
estpost ttest q61 if (q10_treat == 0 | q10_treat == 1), by (q10_treat) uneq
putexcel I5 = matrix(e(b)) J5 = matrix(e(se)) K5 = matrix(e(p)) 

*marriage year
estpost ttest q69 if (q10_treat == 0 | q10_treat == 1), by (q10_treat) uneq
putexcel I6 = matrix(e(b)) J6 = matrix(e(se)) K6 = matrix(e(p)) 

*number daughters
estpost ttest q73 if (q10_treat == 0 | q10_treat == 1), by (q10_treat) uneq
putexcel I7 = matrix(e(b)) J7 = matrix(e(se)) K7 = matrix(e(p)) 

*education
estpost ttest educany if (q10_treat == 0 | q10_treat == 1), by (q10_treat) uneq
putexcel I8 = matrix(e(b)) J8 = matrix(e(se)) K8 = matrix(e(p)) 

estpost ttest educ if (q10_treat == 0 | q10_treat == 1),		by (q10_treat) uneq
putexcel I9 = matrix(e(b)) J9 = matrix(e(se)) K9 = matrix(e(p)) 

*wages
estpost ttest wages if (q10_treat == 0 | q10_treat == 1), by (q10_treat) uneq
putexcel I10 = matrix(e(b)) J10 = matrix(e(se)) K10 = matrix(e(p)) 

*wealth index
estpost ttest wealthindex if (q10_treat == 0 | q10_treat == 1), by (q10_treat) uneq
putexcel I12 = matrix(e(b)) J12 = matrix(e(se)) K12 = matrix(e(p))  

*land title
estpost ttest landt if (q10_treat == 0 | q10_treat == 1), by (q10_treat) uneq
putexcel I13 = matrix(e(b)) J13 = matrix(e(se)) K13 = matrix(e(p))  


				***************************
				***************************
				***************************

				
		
**** Appendix A12
*----------------

*diff-in-diff test w/o vs. w/controls for appendix

* Appendix A12a

eststo: reg q38_outcome i.q0##i.group, robust

eststo: reg q38_outcome i.q0##i.group $controls, robust


esttab using "A12a_final_2020.csv", b(2) se(2) replace ///
		label title(Voter Turnout in Legislative Assembly Elections) ///
		keep(2.q0 1.group 2.q0#1.group) ///
		order(2.q0 1.group 2.q0#1.group) gaps ///
		varlabels(2.q0 "Women" 1.group "Matrilineal" ///
		2.q0#1.group "Women x Mat") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear	
				
	
* Appendix A12b

eststo: reg q34_outcome i.q0##i.group, robust

eststo: reg q34_outcome i.q0##i.group $controls, robust


esttab using "A12b_final_2020.csv", b(2) se(2) replace ///
		label title(Trust in Local Legislators) ///
		keep(2.q0 1.group 2.q0#1.group) ///
		order(2.q0 1.group 2.q0#1.group) gaps ///
		varlabels(2.q0 "Women" 1.group "Matrilineal" ///
		2.q0#1.group "Women x Mat") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear	


* Appendix A12c

eststo: reg q35_outcome i.q0##i.group, robust

eststo: reg q35_outcome i.q0##i.group $controls, robust


esttab using "A12c_final_2020.csv", b(2) se(2) replace ///
		label title(Trust in Political Parties) ///
		keep(2.q0 1.group 2.q0#1.group) ///
		order(2.q0 1.group 2.q0#1.group) gaps ///
		varlabels(2.q0 "Women" 1.group "Matrilineal" ///
		2.q0#1.group "Women x Mat") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear		
	
	
* Appendix A12d

eststo: reg q28_outcome i.q0##i.group, robust

eststo: reg q28_outcome i.q0##i.group $controls, robust


esttab using "A12d_final_2020.csv", b(2) se(2) replace ///
		label title(Perceptions of Local Officials' Accountability) ///
		keep(2.q0 1.group 2.q0#1.group) ///
		order(2.q0 1.group 2.q0#1.group) gaps ///
		varlabels(2.q0 "Women" 1.group "Matrilineal" ///
		2.q0#1.group "Women x Mat") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear			
		
		
				***************************
				***************************
				***************************		
					


* Appendix A13: personal cost treatment (q8), social controls w/ marriage added
*------------------------------------------------------------------------------

fvset base 1 marriage

eststo clear

*pat men

eststo: reg q8_outcome q8_treat  ///
	if (q0 == 1 & group == 0), robust

eststo: reg q8_outcome q8_treat $controlsocpol ///
	if (q0 == 1 & group == 0), robust

	
*pat women

eststo: reg q8_outcome q8_treat ///
	if (q0 == 2 & group == 0), robust

eststo: reg q8_outcome q8_treat $controlsocpol ///
	if (q0 == 2 & group == 0), robust
	
	
*mat men
eststo: reg q8_outcome q8_treat ///
	if (q0 == 1 & group == 1), robust

eststo: reg q8_outcome q8_treat $controlsocpol ///
	if (q0 == 1 & group == 1), robust

	
*mat women
eststo: reg q8_outcome q8_treat  ///
	if (q0 == 2 & group == 1), robust

eststo: reg q8_outcome q8_treat $controlsocpol ///
	if (q0 == 2 & group == 1), robust
	
	
esttab using "A13_final_2020.csv", b(2) se(2) replace ///
		label title(Effect of Personal Cost Treatment on Policy Preferences) ///
		mtitle("Pat Men" "Pat Men" "Pat Women" "Pat Women" ///
		"Mat Men" "Mat Men" "Mat Women" "Mat Women") ///
		keep(q8_treat _cons) order(q8_treat _cons) gaps ///
		varlabels(q8_treat "Explicit cost to policy" _cons ///
		"Constant (control mean)") ///
		indicate("Social Controls = $controlsocpol") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear				
			

				***************************
				***************************
				***************************		
				
					
			
* Appendix A14: Effect of Personal Cost Treatment, Ordered Logit
*---------------------------------------------------------------
eststo clear

* Ordered Logit 
eststo: ologit q8_outcome i.q8_treat##i.q0##i.group, robust
* Ordered Logit controls
eststo: ologit q8_outcome i.q8_treat##i.q0##i.group $controls, robust

esttab using "A14_final_2020.csv", b(2) se(2) replace ///
		label title(Effect of Personal Cost Treatment on Policy Preferences) ///
		mtitle("Logit" "W/Controls") ///
		keep(1.q8_treat 2.q0 1.group 1.q8_treat#2.q0 1.q8_treat#1.group ///
		2.q0#1.group 1.q8_treat#2.q0#1.group) ///
		order(1.q8_treat 2.q0 1.group 1.q8_treat#2.q0 1.q8_treat#1.group ///
		2.q0#1.group 1.q8_treat#2.q0#1.group) gaps ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		indicate("Controls = $controls") ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 

eststo clear


				***************************
				***************************
				***************************	
				
				

* Appendix A16: personal cost treatment (q77), triple interaction
*----------------------------------------------------------------


eststo clear

* Ordered Logit 
eststo: ologit q77_outcome i.q77_treat##i.q0##i.group, robust
* Ordered Logit controls
eststo: ologit q77_outcome i.q77_treat##i.q0##i.group $controls, robust

esttab using "A16_final_2020.csv", b(2) se(2) replace ///
		label title(Effect of Postcard Treatment on Policy Preferences) ///
		mtitle("Ordered Logit" "Ordered Logit Control") ///
		keep(1.q77_treat 2.q0 1.group 1.q77_treat#2.q0 1.q77_treat#1.group ///
		2.q0#1.group 1.q77_treat#2.q0#1.group) ///
		order(1.q77_treat 2.q0 1.group 1.q77_treat#2.q0 1.q77_treat#1.group ///
		2.q0#1.group 1.q77_treat#2.q0#1.group) gaps ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		indicate("Controls = $controls") ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear



				***************************
				***************************
				***************************	

		
	
	
*Appendix A17: Robustness checks 
*-------------------------------

** Does treatment status predict post-card sending?
**-------------------------------------------------


eststo clear


*pat men
eststo: reg q77_outcomesend q77_treat postdist $controls if (q0 == 1 & group == 0), robust
	
*pat women
eststo: reg q77_outcomesend q77_treat postdist $controls if (q0 == 2 & group == 0), robust
		
*mat men
eststo: reg q77_outcomesend q77_treat postdist $controls if (q0 == 1 & group == 1), robust

*mat women
eststo: reg q77_outcomesend q77_treat postdist $controls if (q0 == 2 & group == 1), robust
	
esttab using "A17_final_2020.csv", b(2) se(2) replace ///
		label title(Effect of Treatment on Sending a Postcard) ///
		mtitle("Pat Men" "Pat Women" "Mat Men" "Mat Women") ///
		keep(q77_treat _cons) order(q77_treat _cons) gaps ///
		varlabels(q77_treat "Treatment" _cons ///
		"Constant (control mean)") ///
		indicate("Demographic controls = $controls" ) ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear		



				***************************
				***************************
				***************************	
		
		
	

		
* Appendix A18: Postcard treatment (q77), robust to no controls, social controls
*-------------------------------------------------------------------------------

* NOTE: Our outcome here is policy preferences expressed by postcard senders.

fvset base 1 marriage

eststo clear


*pat men
eststo: reg q77_outcome q77_treat  ///
	if (q0 == 1 & group == 0), robust
	
eststo: reg q77_outcome q77_treat $controlsocpol ///
	if (q0 == 1 & group == 0), robust

*pat women
eststo: reg q77_outcome q77_treat  ///
	if (q0 == 2 & group == 0), robust
	
eststo: reg q77_outcome q77_treat $controlsocpol ///
	if (q0 == 2 & group == 0), robust
	
*mat men
eststo: reg q77_outcome q77_treat  ///
	if (q0 == 1 & group == 1), robust
	
eststo: reg q77_outcome q77_treat $controlsocpol ///
	if (q0 == 1 & group == 1), robust

*mat women
eststo: reg q77_outcome q77_treat ///
	if (q0 == 2 & group == 1), robust
	
eststo: reg q77_outcome q77_treat $controlsocpol ///
	if (q0 == 2 & group == 1), robust
	
	
esttab using "A18_final_2020.csv", b(2) se(2) replace ///
		label title(Effect of Postcard Treatment on Policy Preferences) ///
		mtitle("Pat Men" "Pat Men" "Pat Women" "Pat Women" ///
		"Mat Men" "Mat Men" "Mat Women" "Mat Women") ///
		keep(q77_treat _cons) order(q77_treat _cons) gaps ///
		varlabels(q77_treat "Explicit cost to policy" _cons ///
		"Constant (control mean)") ///
		indicate("Social Controls = $controlsocpol") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear		




				***************************
				***************************
				***************************	
		
		
		
			
	

* Appendix A19 - Heckman Selection Model 
*---------------------------------------

eststo clear


*pat men
eststo: heckman q77_outcome q77_treat if (q0 == 1 & group == 0), select(postdist)
	
*pat women
eststo: heckman q77_outcome q77_treat  if (q0 == 2 & group == 0), select(postdist)
		
*mat men
eststo: heckman q77_outcome q77_treat if (q0 == 1 & group == 1), select(postdist)

*mat women
eststo: heckman q77_outcome q77_treat  if (q0 == 2 & group == 1), select(postdist)
	
esttab using "A19_final_2020.csv", b(2) se(2) replace ///
		label title(Effect of Postcard Treatment on Policy Preferences ///
		(Heckman selection model correction)) ///
		mtitle("Pat Men" "Pat Women" "Mat Men" "Mat Women") ///
		keep(q77_treat postdist _cons) order(q77_treat postdist _cons) gaps ///
		varlabels(q77_treat "Treatment" _cons ///
		"Constant (control mean)") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear	
	
* For rho coefficient values and standard errors, see details in STATA interface

	
	
				***************************
				***************************
				***************************	
			

		
* Appendix A20a: preferences about decision making (q19), with varied controls
*-----------------------------------------------------------------------------

eststo clear

*pat men
eststo: reg q19_outcome q19_treat if (q0 == 1 & group == 0 & ///
		(q19_treat == 1 | q19_treat == 3)), robust
			
eststo: reg q19_outcome q19_treat $controls if (q0 == 1 & group == 0 & ///
		(q19_treat == 1 | q19_treat == 3)), robust
					
*pat women
eststo: reg q19_outcome q19_treat if (q0 == 2 & group == 0 & ///
		(q19_treat == 1 | q19_treat == 3)), robust
		
eststo: reg q19_outcome q19_treat $controls if (q0 == 2 & group == 0 & ///
		(q19_treat == 1 | q19_treat == 3)), robust

*mat men
eststo: reg q19_outcome q19_treat if (q0 == 1 & group == 1 & ///
		(q19_treat == 1 | q19_treat == 3)), robust
		
eststo: reg q19_outcome q19_treat $controls if (q0 == 1 & group == 1 & ///
		(q19_treat == 1 | q19_treat == 3)), robust

*mat women
eststo: reg q19_outcome q19_treat if (q0 == 2 & group == 1 & ///
		(q19_treat == 1 | q19_treat == 3)), robust
		
eststo: reg q19_outcome q19_treat $controls if (q0 == 2 & group == 1 & ///
		(q19_treat == 1 | q19_treat == 3)), robust

	
esttab using "A20A_final_2020.csv", b(2) se(2) replace ///
		label title(Intra-HH decision-making Preferences ///
		Conditional on Income Contribution; baseline = both partners earn) ///
		mtitle("Pat Men" "Pat Men" "Pat Women" "Pat Women" ///
		"Mat Men" "Mat Men" "Mat Women" "Mat Women") ///
		keep(q19_treat _cons) ///
		order(q19_treat _cons) gaps ///
		varlabels(q19_treat "Women earn more" _cons ///
		"Constant (control mean)") ///
		indicate ("Demographic controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear		
		
	

	
				***************************
				***************************
				***************************	
	
			
			

* Appendix A20b: preferences about decision making (q19), triple interaction
*--------------------------------------------------------------------------

* Compare baseline (both make money) with treatment where wife earns more money
eststo clear

* Ordered logit, no controls
eststo: ologit q19_outcome i.q19_treat##i.q0##i.group if ///
(q19_treat == 1 | q19_treat == 3), robust

* Ordered logit, controls
eststo: ologit q19_outcome i.q19_treat##i.q0##i.group $controls if ///
(q19_treat == 1 | q19_treat == 3), robust
* q19_treat#q0#group = .7976301  se = .3677786  p-val = 0.030

esttab using "A20B_final_2020.csv", b(2) se(2) replace ///
label title(Effect of Gendered Wealth Treatment Effect on ///
Preferences, Ordered Logit) ///
mtitle("Ordered Logit" "Ordered Logit Control") ///
keep(3.q19_treat 2.q0 1.group 3.q19_treat#2.q0 3.q19_treat#1.group ///
2.q0#1.group 3.q19_treat#2.q0#1.group) ///
order(3.q19_treat 2.q0 1.group 3.q19_treat#2.q0 3.q19_treat#1.group ///
2.q0#1.group 3.q19_treat#2.q0#1.group) gaps ///
varlabels(3.q19_treat "Wife is main earner" 2.q0 "Female" ///
1.group "Matrilineal Group" 3.q19_treat#2.q0 "Wife is main earner X Female" ///
3.q19_treat#1.group "Wife main X Matrilineal" 2.q0#1.group "Female X Matrilineal" ///
3.q19_treat#2.q0#1.group "Wife main X Fem X Matr") ///
star(* 0.10 ** 0.05 *** 0.01) ///
indicate("Controls = $controls") ///
varwidth(15) modelwidth(10) stats(N,fmt(0) ///
label("Observations"))

eststo clear



		
				***************************
				***************************
				***************************	
				
				
			
				
* Appendix A21a: Intra-household distribution treatment effect with var controls
*-------------------------------------------------------------------------------

* Compare baseline with treatment where head gets more money 

eststo clear


*pat men
eststo: reg q10_outcome q10_treat if (q0 == 1 & group == 0  & ///
	(q10_treat == 0 | q10_treat == 1)), robust
	
eststo: reg q10_outcome q10_treat $controls if (q0 == 1 & group == 0  & ///
	(q10_treat == 0 | q10_treat == 1)), robust
	
*pat women
eststo: reg q10_outcome q10_treat if (q0 == 2 & group == 0 & ///
	(q10_treat == 0 | q10_treat == 1)), robust
	
eststo: reg q10_outcome q10_treat $controls if (q0 == 2 & group == 0 & ///
	(q10_treat == 0 | q10_treat == 1)), robust

*mat men
eststo: reg q10_outcome q10_treat if (q0 == 1 & group == 1  & ///
	(q10_treat == 0 | q10_treat == 1)), robust
	
eststo: reg q10_outcome q10_treat $controls if (q0 == 1 & group == 1  & ///
	(q10_treat == 0 | q10_treat == 1)), robust

*mat women
eststo: reg q10_outcome q10_treat if (q0 == 2 & group == 1 & ///
	(q10_treat == 0 | q10_treat == 1)), robust
	
eststo: reg q10_outcome q10_treat $controls if (q0 == 2 & group == 1 & ///
	(q10_treat == 0 | q10_treat == 1)), robust

	
esttab using "A21A_final_2020.csv", b(2) se(2) replace ///
		label title(Effect of Costly Gov't Distribution on Policy Preferences) ///
		mtitle("Pat Men" "Pat Men" "Pat Women" "Pat Women" ///
		"Mat Men" "Mat Men" "Mat Women" "Mat Women") ///
		keep(q10_treat _cons) order(q10_treat _cons) gaps ///
		varlabels(q10_treat "Costly gov't dist" _cons ///
		"Constant (control mean)") indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear		

	
				***************************
				***************************
				***************************	

				
		

* Appendix A21b:Intra-household distribution treatment (q10), triple interaction
*-------------------------------------------------------------------------------

* Compare baseline with treatment where head gets more money
eststo clear

* Ordered Logit
eststo: ologit q10_outcome i.q10_treat##i.q0##i.group if ///
  (q10_treat == 0 | q10_treat == 1), robust

* Ordered Logit controls
eststo: ologit q10_outcome i.q10_treat##i.q0##i.group $controls if ///
  (q10_treat == 0 | q10_treat == 1), robust


esttab using "A21B_final_2020.csv", b(2) se(2) replace ///
label title(Effect of Gendered Wealth Treatment Effect on ///
Preferences, Ordered Logit) ///
mtitle("Ordered Logit" "Ordered Logit Control") ///
keep(1.q10_treat 2.q0 1.group 1.q10_treat#2.q0 1.q10_treat#1.group ///
2.q0#1.group 1.q10_treat#2.q0#1.group) ///
order(1.q10_treat 2.q0 1.group 1.q10_treat#2.q0 1.q10_treat#1.group ///
2.q0#1.group 1.q10_treat#2.q0#1.group) gaps ///
varlabels(1.q10_treat "Costly gov't distribution'" 2.q0 "Female" ///
1.group "Matrilineal Group" 1.q10_treat#2.q0 "Costly gov't dist X Female" ///
1.q10_treat#1.group "Costly X Matrilineal" 2.q0#1.group "Female X Matrilineal" ///
1.q10_treat#2.q0#1.group "Costly X Fem X Matr") ///
star(* 0.10 ** 0.05 *** 0.01) ///
indicate("Controls = $controls") ///
varwidth(15) modelwidth(10) stats(N,fmt(0) ///
label("Observations"))

eststo clear

		
				***************************
				***************************
				***************************	
									

* Appendix A22: cultural wealth subanalysis, personal cost treat (q8) w/controls
*-------------------------------------------------------------------------------

eststo clear
*pat men - low wealth
eststo: reg q8_outcome q8_treat $controls_alt if (q0 == 1 & group == 0 & ///
		wealthindexaltb == 0), robust
		
	
*pat men - high wealth
eststo: reg q8_outcome q8_treat $controls_alt if (q0 == 1 & group == 0 & ///
		wealthindexaltb == 1), robust
		
		
*pat women - low wealth
eststo: reg q8_outcome q8_treat $controls_alt if (q0 == 2 & group == 0 &  ///
		wealthindexaltb == 0), robust
		

*pat women - high wealth
eststo: reg q8_outcome q8_treat $controls_alt if (q0 == 2 & group == 0 & ///
		wealthindexaltb == 1), robust

		
*mat men - low wealth
eststo: reg q8_outcome q8_treat $controls_alt if (q0 == 1 & group == 1 & ///
		wealthindexaltb == 0), robust
		

* mat men - high wealth
eststo: reg q8_outcome q8_treat $controls_alt if (q0 == 1 & group == 1 & ///
		wealthindexaltb == 1), robust
		

*mat women - low wealth
eststo: reg q8_outcome q8_treat $controls_alt if (q0 == 2 & group == 1 & ///
		wealthindexaltb == 0), robust
		

* mat women - high wealth
eststo: reg q8_outcome q8_treat $controls_alt if (q0 == 2 & group == 1 & ///
		wealthindexaltb == 1), robust

		
esttab using "A22_final_2020.csv", b(2) se(2) replace ///
		label title(Effect of Personal Cost Treatment on Policy Preferences) ///
		mtitle("Pat men P" "Pat men R" "Pat women P" "Pat women R" ///
		"Mat men P" "Mat men R" "Mat women P" "Mat women R") ///
		keep(q8_treat _cons) ///
		order(q8_treat _cons) gaps ///
		varlabels(q8_treat "Treatment" _cons "Constant (control)") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
		
eststo clear		

	
				***************************
				***************************
				***************************	

	
							
*** Appendix A23: Political participation questions w/ and w/o social controls
***---------------------------------------------------------------------------

* Appendix A23a
fvset base 2 q0
fvset base 1 marriage
* pat groups
eststo: reg q38_outcome 1.q0 $controls if (group == 0), robust

eststo: reg q38_outcome 1.q0 $controlsocpol if (group == 0), robust

* mat groups
eststo: reg q38_outcome 1.q0 $controls if (group == 1), robust

eststo: reg q38_outcome 1.q0 $controlsocpol if (group == 1), robust

esttab using "A23a_final_2020.csv", b(2) se(2) replace ///
		label title(Voter Turnout in Legislative Assembly Elections) ///
		keep(1.q0 _cons) ///
		order(1.q0 _cons) gaps ///
		varlabels(1.q0 "Male") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear	
				
	
* Appendix A23b
fvset base 2 q0
fvset base 1 marriage

* pat
eststo: reg q34_outcome 1.q0 $controls if group == 0, robust

eststo: reg q34_outcome 1.q0 $controlsocpol if group == 0, robust
* mat
eststo: reg q34_outcome 1.q0 $controls if group == 1, robust

eststo: reg q34_outcome 1.q0 $controlsocpol if group == 1, robust

esttab using "A23b_final_2020.csv", b(2) se(2) replace ///
		label title(Trust in Local Legislators) ///
		keep(1.q0 _cons) ///
		order(1.q0 _cons) gaps ///
		varlabels(1.q0 "Male") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear	


* Appendix A23c
fvset base 2 q0
fvset base 1 marriage

* pat
eststo: reg q35_outcome 1.q0 $controls if group == 0, robust

eststo: reg q35_outcome 1.q0 $controlsocpol if group == 0, robust
* mat
eststo: reg q35_outcome 1.q0 $controls if group == 1, robust

eststo: reg q35_outcome 1.q0 $controlsocpol if group == 1, robust

esttab using "A23c_final_2020.csv", b(2) se(2) replace ///
		label title(Trust in Political Parties) ///
		keep(1.q0 _cons) ///
		order(1.q0 _cons) gaps ///
		varlabels(1.q0 "Male") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear		
	
	
* Appendix A23d
fvset base 2 q0
fvset base 1 marriage

*pat
eststo: reg q28_outcome 1.q0 $controls if group == 0, robust

eststo: reg q28_outcome 1.q0 $controlsocpol if group == 0, robust
* mat
eststo: reg q35_outcome 1.q0 $controls if group == 1, robust

eststo: reg q35_outcome 1.q0 $controlsocpol if group == 1, robust

esttab using "A23d_final_2020.csv", b(2) se(2) replace ///
		label title(Perceptions of Local Officials' Accountability) ///
		keep(1.q0 _cons) ///
		order(1.q0 _cons) gaps ///
		varlabels(1.q0 "Male") ///
		indicate("Demographic Controls = $controls") ///
		star(* 0.10 ** 0.05 *** 0.01) ///
		varwidth(15) modelwidth(10) stats(N,fmt(0) ///
		label("Observations")) 
eststo clear				
		
	
				***************************
				***************************
				***************************
				
							
* Appendix Table A24a: Support immigration from outside Meghalaya?
*-----------------------------------------------------------------				

* aggregate behavior
putexcel set A24a_immigration_final_2020.xls, replace

*set titles
putexcel C1  = ("Patrilineal") 	E1 = ("Matrilineal") ///
		 A3  = ("Men") 			///
		 A4  = ("Observations") ///
		 A6  = ("Women") 		///
		 A7 = ("Observations")  ///
		 A9 = ("Men - Women")   ///
		 A10 = ("s.e.") 		///
		 A11 = ("p-value")		///
		 A13 = ("Difference (mat-pat)") ///
		 A14 = ("s.e.") 		///
		 A15 = ("p-value")		
		 
		 
* means
		 
* men
*----

*patrilineal
sum q51_support if (q0 == 1 & group == 0)
putexcel C3 = matrix(r(mean)) C4 = matrix(r(N))

*matrilineal
sum q51_support if (q0 == 1 & group == 1)
putexcel E3 = matrix(r(mean)) E4 = matrix(r(N))


* women
*------

*patrilineal
sum q51_support if (q0 == 2 & group == 0)
putexcel C6 = matrix(r(mean)) C7 = matrix(r(N))

*matrilineal
sum q51_support if (q0 == 2 & group == 1)
putexcel E6 = matrix(r(mean)) E7 = matrix(r(N))


* ttests

*patrilineal women vs. men
estpost ttest q51_support if (group == 0), by(q0)
putexcel C9 = matrix(e(b)) C10 = matrix(e(se)) C11 = matrix(e(p))

*matrilineal women vs. men
estpost ttest q51_support if (group == 1), by(q0)
putexcel E9 = matrix(e(b)) E10 = matrix(e(se)) E11 = matrix(e(p))
				

*diff-in-diff test 
* beta = 0.20; p=0.000
eststo: reg q51_support i.q0##i.group, robust
local t = _b[2.q0#1.group]/_se[2.q0#1.group]
local p = ttail(e(df_r), `t')
putexcel C13 = matrix(_b[1.q0#1.group]) C14 = (_se[1.q0#1.group]) C15 = `p'


				
				***************************
				***************************
				***************************

				
* Appendix Table 24b: invite indvl "not of your tribe or community to eat"@home?
*-------------------------------------------------------------------------------

* aggregate behavior
putexcel set A24b_invite_final_2020.xls, replace

*set titles
putexcel C1  = ("Patrilineal") 	E1 = ("Matrilineal") ///
		 A3  = ("Men") 			///
		 A4  = ("Observations") ///
		 A6  = ("Women") 		///
		 A7 = ("Observations")  ///
		 A9 = ("Men - Women")   ///
		 A10 = ("s.e.") 		///
		 A11 = ("p-value")		///
		 A13 = ("Difference (mat-pat)") ///
		 A14 = ("s.e.") 		///
		 A15 = ("p-value")		
		 
		 
* means
		 
* men
*----

*patrilineal
sum q67_invite if (q0 == 1 & group == 0)
putexcel C3 = matrix(r(mean)) C4 = matrix(r(N))

*matrilineal
sum q67_invite if (q0 == 1 & group == 1)
putexcel E3 = matrix(r(mean)) E4 = matrix(r(N))


* women
*------

*patrilineal
sum q67_invite if (q0 == 2 & group == 0)
putexcel C6 = matrix(r(mean)) C7 = matrix(r(N))

*matrilineal
sum q67_invite if (q0 == 2 & group == 1)
putexcel E6 = matrix(r(mean)) E7 = matrix(r(N))


* ttests

*patrilineal women vs. men
estpost ttest q67_invite if (group == 0), by(q0)
putexcel C9 = matrix(e(b)) C10 = matrix(e(se)) C11 = matrix(e(p))

*matrilineal women vs. men
estpost ttest q67_invite if (group == 1), by(q0)
putexcel E9 = matrix(e(b)) E10 = matrix(e(se)) E11 = matrix(e(p))
				
				
*diff-in-diff test 
* beta = 0.20; p=0.000
eststo: reg q67_invite i.q0##i.group, robust
local t = _b[2.q0#1.group]/_se[2.q0#1.group]
local p = ttail(e(df_r), `t')
putexcel C13 = matrix(_b[1.q0#1.group]) C14 = (_se[1.q0#1.group]) C15 = (`p')

				
				***************************
				***************************
				***************************

	
	
* Appendix Table 24c: "invite" varies, matrilineal groups by political activity?
*-------------------------------------------------------------------------------				

* aggregate behavior
putexcel set A24c_invite_vote_final_2020.xls, replace

*set titles
putexcel C1  = ("Didn't Vote") 	E1 = ("Voted") ///
		 A3  = ("Men") 			///
		 A4  = ("Observations") ///
		 A6  = ("Women") 		///
		 A7 = ("Observations")  ///
		 A9 = ("Men - Women")   ///
		 A10 = ("s.e.") 		///
		 A11 = ("p-value")		///
		 A13 = ("Difference (mat-pat)") ///
		 A14 = ("s.e.") 		///
		 A15 = ("p-value")		///
		 A17 = ("Voted-Didn't") ///
		 A18 = ("s.e.") 		///
		 A19 = ("p-value")		 
		 
		 
* means
		 
* men
*----

*non-voter
sum q67_invite if (q0 == 1 & group == 1 & q38_outcome == 0)
putexcel C3 = matrix(r(mean)) C4 = matrix(r(N))

*voter
sum q67_invite if (q0 == 1 & group == 1 & q38_outcome == 1)
putexcel E3 = matrix(r(mean)) E4 = matrix(r(N))


* women
*------

*non-voter
sum q67_invite if (q0 == 2 & group == 1 & q38_outcome == 0)
putexcel C6 = matrix(r(mean)) C7 = matrix(r(N))

*voter
sum q67_invite if (q0 == 2 & group == 1 & q38_outcome == 1)
putexcel E6 = matrix(r(mean)) E7 = matrix(r(N))


* ttests

*non-voter women vs. men
estpost ttest q67_invite if (group == 1 & q38_outcome == 0), by(q0)
putexcel C9 = matrix(e(b)) C10 = matrix(e(se)) C11 = matrix(e(p))

*voter women vs. men
estpost ttest q67_invite if (group == 1 & q38_outcome == 1), by(q0)
putexcel E9 = matrix(e(b)) E10 = matrix(e(se)) E11 = matrix(e(p))

				
*diff-in-diff test 
eststo: reg q67_invite i.q0##i.q38_outcome if group == 1, robust
local t = _b[2.q0#1.q38_outcome]/_se[2.q0#1.q38_outcome]
local p = ttail(e(df_r), `t')
putexcel C13 = matrix(_b[1.q0#1.q38_outcome]) C14 = (_se[1.q0#1.q38_outcome]) C15 = (`p')


				***************************
				***************************
				***************************
				


	
